Since the introduction of callbacks (#533), when using the train function, only a single data point is visible in the TensorBoard logs after training. This occurs regardless of the write_every parameter value, which is intended to control the frequency of intermediate result logging.
Since the introduction of callbacks (#533), when using the
trainfunction, only a single data point is visible in the TensorBoard logs after training. This occurs regardless of thewrite_everyparameter value, which is intended to control the frequency of intermediate result logging.